Proposal for reV1Sl0n of ISO 4377 Flat-V weirs
J.P.Th. Kalkwijk Wang Lianxiang
I
LaboratoryDepartment oof Civil Engineeringf Fluid MechanicsI
I
I
I'
I
'
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
Proposal for revision of ISO 4377
Flat-'V' weirs
J.P.Th. Kalkwijk
Wang Lianxiang
Report no. 19 - 83
Laboratory of Fluid Mechanics
Department of Civil Engineering
Delft University o
f
Technology
I
I
I
I
Elucidation to the Netherlands' proposal for revision of ISO 4377
,
Liquid flow measurements in open channels
Flat-V weirs
I
I
In ISO/TC 113/SC 2/308 it was suggested that ISO 4377 could be
significantly simplified as far as the calculation of discharge
was involved.
The proposal for revision is based upon this
suggestion.
The revision ~s fully bàsed upon the material given in ISO 4377
.
In that respect no changes have been made.
The result of the revision is~ that
i
the calculation of discharge is simple and straightforward
ii
the number of tables (sometimes very extensive) has reduced
from 13 to only 5.
I
I
I
I
I
I
I
I
I
I
I
I·
I
I
.
.I
I
I
I
7.
7. 1I
I
1
I
I
I
·
I
7.2
I
I
I
I
I
I
7.3
I
I
I
I
I
..
-1-Discharge equation EquationThe equation of discharge ~s based on the use of gauged head:
5/
1/
.
IJ
5/
Q
=
(t)
2(t)
2 CDCvCSCdr m g 2h' 2 (1),where
Q
is the dischargeCD is the discharge coefficient
C is the velocity of approach factor
v
Cs is the shape factor
Cdr is tbe drowned flow reduction factor
m is tbe crest cross-slope
g ~s the acceleration due to gravity
h is the upstream gauged head relative to the lowest crest
elevation.
Discharge coefficient
The discharge coefficient 1S g~ven by
CD
=
CDm<hè/h)5/3=
CDm(1 - km
/h)5/2 (2)where
CDm is the modular coefficient of discharge,
~~ is the effective gaugeg head, which is equal to (h-km),
k is the he ad correction factor~
m
Both CUm and ~ depend on the cross-slope of the weir. They are
given in table 3.
In the case of field measurement, km ~s negligible because it ~s
less than 1 mmo
Velocity of approach factor
Since the discharge equation ~s d~rived from a formula in which
the total head is used, a Vèlocity of approach factor is introduced
to correct the gauged head. That is
C
=
(R/h) 5l :
v (3)
where
R is the upstream total head with respect to the lowest crest
elevation.
I
,
I
I
I
I
I
I
I
'
1
I
I
I
7.4I
I
I
I
I
I
7.5
I
I
-2-(4), wherePI
1S the difference between the lowest crest elevation and themean bed level in the approach channel.
b is the crest width
r
(5)Equation (4) can be solved iteratively; the values of C; are
v
obtained in terms of YI and given in table 4.
In the case of
Y
I <0,08
(Cv-I)« I, a good approximation forCV 1S given below
C v
1, 2SY I
=
I +-=--=-=-
1-2,SYI
(6)The relative error due to this approximation 1S less than 0,7 7.
Shape factor
A
shape factor is introduced into the discharge equation forflat-V weir because the geometry of flow changes when the
discharge exceeds the V-full condition. Thus:
if
h
e_<h'
(8)
if h, > h'
e
where h' = h/2m is the difference between the lowest and highest
crest elevations.
Drowned flow reduction factor
When the weir gets drowned, i.e., h >0,4He, the discharge decreases.
pe
I
.~.-3
-1
1
Cdr := C .=
dr if h /R < 0,4 pe e-1,078j-0,909-ChI
H
.
)3/2J
0,I8_3ifh /H > 0,4 _ pe e pe e (9) where·
1
hpe is the effective separation pocket head relative to the lowest crest
elevation, h :=h - k •
pe p m
i$ the gauged separation pocket head relative to the lowest crest
elevation.
is the effective upstream total head relative to the lowest
2
crest elevation,
H
:=h + v /2g - k •e m
I
1
h p H eI
I
Raving solved equation (9), the values of Cdr are obtained and
given in Table 5, in terms of two parameters.
Y2(:= cDCsmh2/bCPl+h).
h [t«, and
pe e
1
In case of hpe/he < 0,9 Cdr can be calculated by us~ng two-stepprocedure with a good approximation (relative error <
1
%
).
Thismethod is as follows:
I
I
a) To estimate the value of CdrO according to the value of h /h.,pe e CdrO 0,I if h /h < 0,55 pe e CdrO := 0,9 if 0,55 < h /h. < 0,70 - pe e (J0) Cdro := 0,8 if 0,70 < h /h < 0,85 - pe e Cdro :=0,75 if 0,85 < h /he < 0,90 - pe
I
I
I
b) To calculate the following quantities,
CDCSmh2 .. Y 2 :=b(PI+h)
I
I
I
Y1 :=0,8 cIro y2 2 CO,4 I + 0,5 2 (1 I) := YIC v v He :=COv,4 he I,078~, 909J
O•183 Cdr := - (hpe/Re)3/2I
I
'When h Ihe> 0,9, more steps of iteration are needed and table 5
pe
-is recorrrrnended.
I
I
I
7.6
7.6.1
I
I
I
7.6.2
I
:1
I
I
I
I
I
I
I
I
I
I
I
I
-4-Limits of app1icationThe practical 10wer limit of upstream head is re1ated to the magnitude of the inf1uence of f1uid properties and boundary roughness. For a we11-maintained weir with a smooth crest section, the minimum head recommended is 0,03 m. If the crest is of smooth concrete or a material of simi1ar texture, a 10wer limit of 0,06-m is suggested.
There is a1so a 1imiting va1ue for the ratio h'/PI of
2,5
and t.here are limitations on HdP2 as shown in t'able 3. These are governed by the scope of experi~ta1 verification and vary with cross-s10pe. P2 is the e1evation of the 10west crest e1evation relative to the downstream bed level.I
-5-I
I
8.
8. 1 Computation of discharge Computation stepsEquation(l) is used for computing discharges from the known
quantities •. Pltm,h' ,b,h and h. The calculation proceeds as
p
I
I
fola) lows:Calculate h "= h - k and CD using equation (2) and thee m
appropriate va lues of CD and k from table 3.
m m
Calculate Cs using equation
(8).
~alculate Cdr. If it is in modular flow comditions, Cdr=)·
"If it is in drowned flow conditions, calculate h =h-k
pe p m
of k from table 3; calculate
m Y2 = CDCS
mh
2/b(Pj+hY; look up using the..
I
b) c)I
I
the ratio appropriate value h /h; calculate pe eI
the appropriate value of Cdr from table 5, or calculate Cdr
according to equations (10) and (11)~
d) Calculate Yl uS1ng equation (5); Determine the value of C by
v
I
e) Determineusing equationthe value of(6) or table 4.Q
by substituting the known and thecalculated values into equation (1).
Examples are given in clause 10.
I
I
8.2
Accuracy8.2.1
The overall accuracy of measurement wil1 depend on:a) The accuracy of construction and finish of the weir;
b) the accuracy of the head measurements;
c) the accuracy of other measured dimensions;
d) the accuracy of the coefficient values;
e) the accuracy of the form of the discharge equation.
I
I
I
I
The method by which estimates of these constituent uncertainties
may be combined to give the overall uncertainty 1n computed discharges
is given in clause 9.
I
8.4.2
The uncertainties (95% confidence limits) on modular dischargecoefficients are given in table 3. These reflect the random and
systematic errors which occur 1n calibration experiments and also
the real but marginal changes 1n coefficient values which occur
changing discharge.
I
I
I
I
I
-6-I
I
9. Errors in flow measurementI
9.1
General9.1.1
The uncertainty of any flow measurement can be estimated if theuncertainties from various sourced are combined. These contributions
to the total uncertainty may be assessed and will indicat.e whether
the rate of flow can be measured with sufficient accuracy for the
purpose in hand. This clause prbvides sufficient information for
estimating the uncertainties of measurements of discharge.
I
I
I
9.1.2
The error in a result is the difference hetween.the true rate offlow and that calculated using the discharge equations quoted in
this International Standard. Thus the error is, by definition,
unknown.but the uncertainty of the measurement may be estimated.
The term uncertainty denotes the deviation from the true rate of
flow with which the measured rate of flow is expected to lie some
nineteen times out of twenty (the 957. confidence limits).
I
I
~.-".'-""- - ~.9.2
Sources of errorI
9.2.)
The sourees of error in the discharge measurement can be identifiedbyeons idering the form of equa tion (1) i.e ,
I
(1)I
The constant(t)5/2 <t)1/2
~s not subject to error and error ~n gmay be ignored. Hence the sources of error which need to he considered
I
are:a) The discharge coefficient CD which essentially has the same value
of uncertainty as C
Dm'
given in tahle 3.The velocity of approach factor C. The following approximate
v
expression may be used to determine the uncertainty in C •
v
I
b)I
Xc
=
0,5 hiP)
(%)
v(12)
I
c) The shape factor cS.uncertainty in this value.When heWhen2
h',he >Cs
h', the= 1 and there ~s novalue ofCs
depends on h' and h , see equation (8). Both these quantities
e
will he of reasonahle magnitude and errors ~n heads will not
normally be significant at this stage. Thus the uncertainty
~n
Cs
~s negligible, i.e. XCs = o.
I
I
I
I
-1-I
I
d) The.crest cross-slope, m. Nlimerical values will dep end onthe 'accuracY'of construction and subsequent measurement of the
structure.
I
e) The upstream gauged head, h. The uncertainty 1n h depends on
uncertainties in head measurement, zeroing of the gauge and
uncertainties associated with the number of readings. Thus
lOOVeh2+eh 2+(2Sh-)2' o (%) h
=
+ (13)I
I
whei1eI
f)eh is the uncertainty in the measurement of upstream head.
Any uncertainty which does not change randomly during a series
of measurements should be included here for example backlash
and friction;
eh is the uncertainty in the determination of the gauge zero;
'0
2Sh
is the uncertainty in the mean of n readings of upstreamhead, see 9.3. It is associated with the random fluctuation 1n
a series of measurements.
The uncertainties eh and e~ dep end on an assessment of probable
uncertainties by the user.
The separation pocket head, h The uncertainty in hp depends
p
on uncertainties in head measurement, zeroing of the gauge and
uncertainties assosiated with the number of readings.
Thus
I
I
I
I
I
I
IOO~h + eh + (2Sh
)2~
;
= .:_
--_.!,..P---,h-p__:.o----fl---{%) (14)I
I
whereeh 1S the uncertainty in the measurement of the separation
p
pocket head. The sources of errors are systematic, for example
backlash and friction;
eh is the uncertainty 1n the determination of the gauge zero;
o
2Shp is the uncertainty in the mean of n readings of the
separation pocket head, see 9.3. It is associated with the
random fluctuatl.ons in a series of measurements.
The uncertainties eh and eh depend on an estimate of probable
p 0
errors by the user. If only one reading of head is made, then
the random uncertainties 2S- or 2S- must be estimated, see 9.5.4.
h hp
I
I
I
I
I
I
-8-I
I
g) The drowned flow reduction factor, Cdr' There are three factorswhich influence the uncertainty 1n Cdr: the uncertainty iri the
laboratory determination of the Cd versus h /He·relationship;
r pe
·The uncertainty in the measurement of the upstream head, h;
and the uncertainty in the measurement of the separation. pocket
head, hp •
I
I
I
A suitable expression for the combined uncertaintyis :
"
I
(15)
I
9.3
Random and systematic errorsErrors can be classified as random or systematic, the formeraffecting the reproducibility of measurement and the lat ter affecting
its true accuracy.
I
I
9.3.1
The standard deviation of a set of n measurements of a variabieR may be estimated from the equation
I
~
n -
~lh
S =
L
(y - y)YIn - 1 (16)
1
where y is the observed mean.is then given by
The standard deviation of the mean
I
Ss-
= J_y
Ii..
(J 7)
I
I
and the uncertainty of the mean 1S2S-
y (for 95% probability) ifthe number of readings, n, is large.
I
I
·
9.3.2
A measurement can also be subject to systematic errors and themean of a large number of measured values would still differ from
the true value of the quantity being measured. An error in a
gauge zero, for example, will produce a systematic error. As
repetition of the measurements does not eliminate systematic
errors, the actual value could only be determined by an independent
measurement known to be more accurate.
I
I
I
I
I
-9-I
I
9.4
Errors in quantities given in this International Standard9.4.1
All the errors in this category are systematic. The values of the discharge coefficients, etc., quoted in this International Standard are .based on an appraisal of experiemnts, carefully carried out with sufficient repetition of readings.I
I
I
9.4.2
However, when measurements are made on other similar installation, systematic discrepancies between coefficients of discharge may occur due to variation in the surface finish of the device, its installation, the appr9ach flow conditions, etc •.I
I
I
9.4.3
The probable uncertainties in the coefficients quoted in previous clauses of this International Standard are based on a consideration of the deviation of experimental data from the given workingequations and a comparison of the equations themselves.
I
9.5
Errors in quantities measured by the user.I
9.5.J
Both random and systematic errors will occur in measurements 1n this category.I
9.5.2
Since neither the methods of measurement nor the way in which theyare to be made are specified, no numerical values for uncertainties in this catègory can be given.
I
I
9.5.3
The uncertainties in the gauged head should be determined from an assessment of the separate soureed of uncertainty, for example the gauge sensitivity, the zero uncertainty, temperature effects, the backlash in the indicating mechanism, the residual random uncertainty in the mean of a series of measurements, etc.I
I
I
I
~
9.5.4
The above component uncertainties should be calculated as percentage standard deviations at the95%
confidence limits but when the value of the component uricertainty 1S determined from only a singlemeasurement, the uncertainty 1S said to be rectangular distributed and may be taken, for the purposes of this International Standard, the limits (plus or minus) within the true value is known to lie with certainty (i.e. half the estimated maximum deviation).
I
I
I
I
ft-10-I
I
9.6
Combination of uncertainties to gl.ve the overall uncertainty in discharge9.6.1
The uncertainty in discharge is given by the expression+
v'
X 2 1 .XQ
=
- C + X2 + X2 + X2 + 6.25x{
(l8) Cv Cdr m DI
I
I
where XQ is the uncertainty.in the ccimputed discharge (per cent).I
I
9.6.2 It should be noted that the uncertainty in discharge is not single valued for a given device, but varies with flow. It may therefore be necessary to consider the uncertainty at several discharges covering the required range of measurement.
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
-11-I
10.
Examples
10.1
Modular flow at low discharge (he
<hl)
lö.i;] À
flat-V weir has a crest cross-slope of 1 : 20,3.
The crest
width and approach channel width are both 36 m an9 the mean upstream
bed level is 0,82 m below the lowest
·
crest deviation.
Calculate the discharge when the observed upstream gauged head is
0,62 m.
Ten successive readings of this head prod~ce
·
a standard
deviation of the mean of 0,5
mmand the estimated uncertainty in
the gauge zero ~s 1
mmoThe basic measurements with their estimated
uncertainties are given below:
I
I
I
m
=
20,30
(.:!:_0,2
i. )b
=
36,00 m
(+0,005 m)
p~=
0,82 m
(.:!:_0,00] m)
h
=
0,621 m
(.:!:_0,003 m)
hl
=
0,887 m
(+0,001 m)
eh
=
.:!:. 0,001 m
02S-
=
t0,001 m
h
:""
.
r>••I
I
I
I
The appropriate coefficient and head correction values are obtained
from table 3, as follows:
CDm
=
1,22
(.:!:_3,2
i.)k
=
0,0005 m
(.:!:_0,0002 m)
mI
10.1.2 Calculation of discharge
(See8.1)
a)he
=h - k
=
0,6205 m
m
5/
CD
CDm(he/h)
2 =0,6188
Cs
=1,0
I
I
b)I
c) d)I
I
I
I
I
=
0,0056
C
=
+1,25YL
=
1,0071
v l"'72;)Y1
5/ 1/ 1/2 5/2 = (%) 2 (~)2 CDCvCSCdr m
g
h
e)Q
3
=
9,63 m /s
3I
1
-12-1
10.1.3 Uncertainty in calculated discharge
(See 9.2)
]0.2
Drowned flow at high discharge
10.2.1 A flat-V weir has a crest cross-slope of 1 : 10.1.
The crest
width and the approach channel width are both 25 m, and the mean
upstream bed level is 0,56 m below the lowest crest elevation.
Calculate the discharge when the upstream and crest tappings record
heads of 2,614 mand
2,211 m respectively.
Five successive readings
of the upstream head produc~d a standard deviation of the mean
of 1,5
mmand the estimated uncertainty in the gauge zero is 2
mmoFive successive readings of the separation pocket head produce a
standard deviation of the mean of 2,1 mm and the estimated uncertainty
in the gauge zero is also 2
mmoThe basic measurements with their
estimated uncertainties are g~ven below:
1
1
1
I
I
1
I
1
I
1
1
1
1
1
I
·
1
I
I
a)From table 3 Xc
=+ 3,2 %.
DFrom equation (]2)
Xc
= ~0,5 h/P1
(%)
v= ~
0,38 %
.: b) c) d) e)For modular flow X
Cdr
= +0,2 %
=o.
From data X
.
mFrom equation (13)
100~h2+eh~+(2Sh)2
~ =
+ h(%)
~ ~ 0,53 %
f)
By using equation (18), we have
X
Q
=
+
I
X 2 + X 2 + X 2
.
+ X2 + 6 25 ~2
-
CD
Cv
Cdr
m
'
-n
=
.:!:.
3,42 %
Thus the uncertainty in discharge (95% confidence limits)
is
.:!:.
3,42 %.
m
'
=
10,1
(.:!:.
0,2 %)
b
=
25,00 m
(~ 0,004 m)
PI=
0,56 m
(+ 0,002 m)
h
=
2,614 m
(_!0,003 m)
hp
=
2,211 m
(~ 0,003 m)
I
I
-13-I
h'
eh
o2sii
2S-hp
=
1,238
mC+
0,001
m)=
.!.0,002
m = +0,003 m
=
+0,0042
m,I
I
The appropriate coefficient and head correction values are obtained
from table 3,
as follows:
I
I
C
Dm=
1,22
k '=0,0008
m mc.:!:.
2,3
%) C.!.0,0002
m)I
I
10.2.2 Calculation of discharge
(See
8.1)a)
h
·
=
h -
k=
2,6132 m,
say 2,613 m
e mCD
=
C
DmChe./h)512
=
0,6194
t", "I
I
b) Cs=
I - C1 - ~ ) 512he
=
0,7991
,g:.) h'_"';;'h - k =2,2102
msay
2,210
m,pe
_
p mh
lhe
=
0,8458
pe
Y2
=
C
DCS
mh2jbCP
i
+h)
=
0,4304
from table 5, Cdr
=0,7880 Cinterpolated)
I
I
I
d),I
I
=
0,0736
C
= v +1,25Y
I1-2,5Y
1=
1,1127I
e)I
I
I
=
121,35
m3jsThus th
e
co
m
puted dischar
g
e
is
121,35
m2js.I
I
-14-I
10.2.3 Uncertainty in calculated discharge
a)
From table 3
Xc
= ~
2,3 %
D
b) Xc
= ~
0,5 h/Pl
(%)
v
=
2,33 %
.
c)
From drowned flow
I
I
I
=
+- 0,18
%
100/eh
2
+ehz+(
2
S- )z
=
+ p 0 hp hp= ~
0,25 %
(%)
From equation (15)
Xc = ~ 5 (1 -Cdr)
ft
+ ~ +~p
dr
(%)
=~1,11%
I
d)
I
I
I
e)X
=
+ /X
2+ X
2 +X 2
+ X
2+6,25
x.
2 Q - CD Cv Cdr . m"n
= ~
vi
2,3
2 +2,3
2+1,11
2+0,2
2+6,25
X0,18
2= ~ 3,49 %
I
Thus the uncertainty in the calculated discharge (95% confidence
~
limits)is~ 3,49
x.
I
I
I
I
I
I
I
I
I
I
AbI
BCI
CDCDm
CdrI
Cs C vI
eh eh 0I
hgI
HH max hI
h'p k mI
KI,KZ mI
n PI
Qs-I
vhI
Xc XD CvXc
drI
~I
XY1QI
YZI
I
Symbols and dimensions
Area of cross-section of flow
Crest width
Width of approach chennel
.Coefficient
Dischar.ge coefficient
Modular coefficient of discharge Drowned flow reduction factor Shape factor
Velocity of approach factor
Uncertainty in the head measurement
Uncertainty in the gauge zero
Acceleration due to gravity
Gauged head above lowest crest deviation Total head above lowest crest deviation
non-dimensional non-dimensional non-dimensional non-dimensional non-dimensional non-dimensional L L
L/tZ
L L Maximum upstream total head above lowest crest elevation LSeparation pocket head L
Difference between lowest and highest crest elevations L
Head correction factor L
non-dimensional Constants
Crest cross-slope (1 vertical /m.horizontal)
Number of measurements in a set
Difference between mean bed level and lowest crest
elevation L
Discharge L3/t
non-dimensional
Standard deviation of mean of several head readings L
Mean velocity at cross-section L/t
Percentage uncertainty 1n discharge coefficient non-dimensional
Percentage uncertainty 1n velocity factor non-dimensional
Percentage uncertainty in drowned flow reduction
factor
Percentage uncertainty 1n head measurement
Percentage uncertainty in discharge measurement
Parameter 1n calculation of velocity factor
Parameter 1n calculation of drowned flow reduction
factor
Coriolis energy coefficient
non-dimensional non-dimensional non-dimensional non-dimensional non-dimensional non-dimensional
I
I
I
1
I
I
I
I
I
·
1
I
I
I
1
I
I
I
I
I
I
I
Subscript 2denotes upstream values
denotes downstream values
denotes "effective" and i:mplies that corrections for fluid
effects have been made to the quantity.
I
I
I
I
I
I
... J•••I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
Table 3. Summary of recommended coefficients., limitations and uncertainties
FLAT-V weir
1.
R/h' <1,0
Modular coefficient, CDm
x
Read correction factor,Km
Uncertainty in coefficient,Xc Dm Modular Limit Other limitations:h'/Pl
h'lP
2 Upstream tapping Crest cross-slope1 : 10
1 : 20
1:40
or less _.. ~1,23
1,22
.-
1,21
0,0004
m0,0005
m0,0008
m +3,0 %
~+3,2 %
.!.
2,9 %
-
-6S
to7S %
6S
to7S %
65
to75 %
-""...•-
~...__.,~2,5
~2,5
~2,5
<2,5
~2,5
~2,5
-10
h'10
h'10
h' 2. R/h' >1,0
Modular coefficient, Cx
Dm Read correction factor,Km
Uncertainty in coefficient, XcDm Modular limi t Other limitations:
h'/P(
h'lp
2 Upstream tapping1
·
,24
1,23
1,22
0,0004
m0,0005
m0,0008
m.!.
2,5 %
.!.
2,8 %
.!.
2,3
%65
to75
%65
to75 %
65
to75
% <2,5
<2,5
<2,5
-
-
-<8,2
<8,2
<4,2
-
-
-10 h'
10 h'
10 h'
;::Computations under non~modular conditions should be based on CDm
=
1,25
,
1,24
and1,22
respectively.I
I
I
I
Y1 0,000 0,002 0,004 0,006 0,008 0,00 1,000 1,002 1,005 1,008 1,010 0,01 1,°
13 1,016 1,018 1,021 1,.024 0,02 1,027 1,030 1,032 1,035 1,038 0,03 1,041 1,044 . 1,047 1,050 1,054 0,04 1,057 1,060 1,063 1,067 1,070 0,05 1,074 1,077 1,080 1,084 1,088 0,06 1,092 1,096 1,100 1,104 1, 108 0,07 1,112 1,116 1, 120 1,124 1,129 0,08 1,133 1,138 1,143 1,148 1,153 0,09 1,158 1,163 1,168 1,174 1,180 0,10 1. ]85 1,191 ] , 197 '1,203 ] ,2]°
0,11 ] ,216 1,223 1,230 1,237 1,244 0,12 1,252 1,260 1,268 1,277 1,286 0,13 1,295 1,305 1,316 1,326 1,338 0,14 1,350 1,363 1,377 1,392 1,408 0,15 1,426 1,446 1,468 1,492 1,523I
I
.
.
I
I
I
I
I
·
I
I
Table 4 Velocity of approach coefficient in terms of Y1